Author:  Sun, Yingjie 
Title:  Studies of slope stability problems by LEM, SRM and DEM 
Degree:  Ph.D. 
Year:  2013 
Subject:  Slopes (Soil mechanics) Hong Kong Polytechnic University  Dissertations 
Department:  Dept. of Civil and Environmental Engineering 
Pages:  xi, 203 leaves : ill. (some col.) ; 30 cm. 
Language:  English 
InnoPac Record:  http://library.polyu.edu.hk/record=b2653051 
URI:  http://theses.lib.polyu.edu.hk/handle/200/7256 
Abstract:  Slope stability problem is a major problem in geotechnical engineering with influence on structure and human life, and slope stability problem has drawn the attentions of many researchers and engineers for the past several decades. This study is aimed to investigate slope stability problem with a better understanding of the failure mechanism and some fundamental principles in slope stability analysis by several methods so that the complete stability and failure processes are investigated. From the present study, some outstanding fundamental questions in slope stability problem have been settled, and the works are beneficial to both academic and practical aspects. This study first begins with the typical upper bound limit equilibrium analysis where different modern heuristic optimization algorithms are modified and improved to locate the critical slip surface efficiently and precisely. This problem has been studied by many researchers in the past, but there is a major difficulty in this problem in that the objective function is nonsmooth and nonconvex and the solution might be trapped into local minimum easily. Towards this complicated problems, two modified optimization algorithms: improved harmony search method MHS and coupled algorithm of HS/PSO are developed. These two algorithms are demonstrated to be more efficient than the original methods, and are particularly suitable for highly complicated problems where there are several strong local minima in the solution domain. The knowledge and works gained in this part of work are useful for practical engineering and also become part of the tools for the later sections. Secondly, the extremum principle and the concept of variable factor of safety based on Pan's postulate and equivalent variational principle are developed in this study. Using the new concept which can be viewed as an equivalent lower bound method, the long outstanding question on the interslice force function is finally settled using the mathematical tool developed in the first part of the present study. Slope stability problem can now become a statically determinate problem, and the interslice force function is actually taken as a variable instead of a prescribed function. This function is now determined by an equivalent lower bound principle which is missing in all the previous limit equilibrium formulation, which is major breakthrough in the basic formulation of the limit equilibrium method. Besides the new extremum LEM formulation, the author has also employed SRM to study the interslice force function. In general, it is found that the interslice force function is close to a bell shape and is also in agreement with the results from LEM, and such results clearly demonstrates that this function cannot be arbitrarily specified as what has been done for more than 40 years. The location of the thrust line also agrees well with the Janbu's Rigorous method which is at 1/3 of slice height from the base for normal cases. As a further extension of the works, the extremum formulations are further extended to the concept of variable factor of safety formulation which can satisfy all the global and local equilibrium. Using this new concept, the stress redistribution and residual strength concept can be cast into the LEM framework under a rigorous lower bound formulation. Progressive failure can now be cast into the framework of limit equilibrium method which is not possible in the past. The limitation of both LEM and SRM is the requirement on continuity which is not possible after the initiation of failure. The failure and postfailure mechanism are investigated by the use of Distinct Element Method (DEM) due to the demand in the consideration of large scale postfailure deformation. The use of DEM to investigate the slip surface is seldom considered in the past but has been achieved in the present study. The effect of water seepage on slope stability using a DEM approach is also an outstanding work which is worth to be investigated. In this study, it is found that the geometry of slope changes continuously, and tensile failure at the crest and shear failure in the middle of the slope are found. The failure mode for soil nailed slope and slope with by water flow are also studied by the DEM with interesting results obtained. For threedimensional problems, there are several interesting problems to be considered. Threedimensional effect of curvature with different nailing modes is considered by SRM, and the intercolumn force function is investigated (which is an outstanding item up to present). It is found that the intercolumn force function within the principal section containing the sliding direction is dominating over other sections. Concave geometry also gives higher global stability which is important for many highway slopes. For nailing pattern, the radial nailing mode gives lower factor of safety for convex slope but higher factor of safety for concave slope as compared with the parallel nailing mode. These results are both useful to the engineers as well as to the basic understanding of threedimensional slope stability problem. Based on the above research involving different methods, many fundamental principles and outstanding problems in slope stability analysis have been settled in the present research. For example, the search for critical failure surface can now be carried out with very high level of confidence even for very complicated problem. Many engineers arbitrarily assign interslice force function (f(x)) equal to 1.0 (or sine function) without any thought, as all textbooks and research papers give the view that this function is "fundamentally indeterminate" and is not critical for normal condition. The author has however pointed out the mistake of this common belief accepted by engineers/researchers for more than 40 years, and has also demonstrated that there are also cases where f(x) is important and has proposed a systematic way to determine this function for arbitrary problem based on the equivalent lower bound principle. This work is then extended to threedimensional condition where no one has ever proposed any interslice force function, and this threedimensional function can now be treated as determinate function. The knowledge about the initiation and post failure movement of slope by DEM has also provided clearer picture about the movement and internal stress distribution within a slope at different stages which is useful to assess the postfailure behaviour and the precautions that are required. 
Files  Size  Format 

b2653051x.pdf  3.838Mb 


As a bona fide Library user, I declare that:  


By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms. 